A Robust Watermarking Technique Based on Image Dual Interpolations

The image interpolation technique is usually used in resampling image resolutions and the relationship between estimated data and adopted interpolation is inseparably. That is, if the interpolation kernel is in a linear mode, the estimated data will process linear characteristic. Similarly, if the estimated data processes non-linear characteristic, the interpolation kernel also is in a non-linear mode. By applying this characteristic, we propose a spatial robust watermarking technique which applies two different interpolation techniques. The statistical correlation coefficient is used to justify which interpolation technique is used to embed watermark signal. Experimental results reveal good visual quality of the embedded image and high robust in various attacks.